Transcript Clustering

Out of Specification Results
(OOS)
A One Day Workshop
Presented by: Karen S. Ginsbury
PCI Pharmaceutical Consulting Israel Ltd.
For IFF
March 2007
Purpose of Workshop

Understand Current Industry Practice as
it relates to all aspects of handling:
– Out of Specification
– Unusual
– Out of Trend Results

When is retesting legitimate
 When does testing have to stop
 When do the authorities need to be notified
PCI Pharmaceutical Consulting Israel Ltd
2/82
Workshop Structure
09:00 – 10:30
 10:30 – 11:00
 11:00 – 12:30

12:30 – 13:30
 13:30 – 15:00

15:00 – 15:15
 15:15 – 16:30

Background and Overview: Barr and Able
Coffee Break
The Making of the FDA Guide
Case study: what NOT to do
Laboratory investigation and checklist
Lunch Break
FDA Guide Continued:
Extended investigation, retesting protocol
Qualitative and quantitative tests
Case study
Tea Break
Case Study and workshop wrap-up
Reporting results; reporting to regulators
PCI Pharmaceutical Consulting Israel Ltd
3/82
PART 1 – BARR and ABLE

Litigation
 USA regulatory environment around
late 1980’s through 1993
 Current regulatory environment in US
 Able - 2005
PCI Pharmaceutical Consulting Israel Ltd
4/82
Some Definitions

What is an OOS Result?

What is a test result

Definition of reportable value
“A reportable value is the end result of the complete measurement
method as documented. It is the value compared with the
specification, the values collected when the term replicates is used,
the values used for official reports, and the values used for any
statistical calculation or analysis.”
 Torbeck (February 1999, Pharmaceutical Technology)

PCI Pharmaceutical Consulting Israel Ltd
5/82
Before Barr – Current Practice



Prior to 1993 and the court decision – it was
COMMON practice to retest
once
or in exceptionally good companies twice
and to release the batch if the retest result was
within the specification
Companies had not really thought about the
practice
But then…nor had the regulators
PCI Pharmaceutical Consulting Israel Ltd
6/82
The Barr Court Case and Judge Wolin
From the New York Times
February 6, 1993, Saturday
(AP); Financial Desk
COMPANY NEWS; Judge Rules On Barr Labs
A generic drug manufacturer must recall batches of
some of its medicines and stop distributing
others until the company completes studies of its
manufacturing process, a Federal judge ruled on
Thursday. But United States District Judge Alfred M.
Wolin refused a request by Federal pharmaceutical
regulators to order a complete shutdown
PCI Pharmaceutical Consulting Israel Ltd
7/82
Barr and OOS




Faced with potential closure, the company
took FDA to court
The judge went into great details as to the
meaning and implications of OOS results
The outcome: FDA draft guidance: 1998
FDA final guidance: 2006
PCI Pharmaceutical Consulting Israel Ltd
8/82
Barr: What happened in court



The judge heard experts on
behalf of FDA and Barr regarding
the practice of retesting
FDA wanted retesting to be banned
under all circumstances
After a long hearing at which five industry experts, an
FDA investigator, and several company employees
testified, Judge Alfred M. Wolin, U.S. District Judge for
the District of New Jersey, issued a 79-page opinion
PCI Pharmaceutical Consulting Israel Ltd
9/82
The Barr Court Case 1993

Reported problems include
– misplaced records
– test data recorded on scrap paper
– failure to control manufacturing steps such as
those governing products' physical properties
– release of products not meeting their
specifications
– inadequate investigation of failed products
PCI Pharmaceutical Consulting Israel Ltd
10/82
Barr: “Testing into Compliance”


Barr had numerous failures
Performed retests with
– no investigations
– no regard for process and product history


Tested until results met specifications
Then irrespective of previous OOS results
for the batch, released product
reporting only the passing results
Q: How do you report passing OOS’s on COA?
PCI Pharmaceutical Consulting Israel Ltd
11/82
Reading the Judgment




Reading the Barr Court Judgment is like
reading FDA’s draft guidance and pretty similar to the
final guidance
Judge Wolin preferred to use the term
"out-of-specification" (OOS) laboratory results rather
than the term "product failure" which was more
common to (preferred by?) FDA's investigators
Ruled that an OOS result identified as laboratory error
by a failure investigation or an outlier test, or
overcome by retesting is not a product failure
BUT
Limited situations where laboratory error could be used
PCI Pharmaceutical Consulting Israel Ltd
12/82
Guide to Inspection: QC Labs

Issued July 1993 (must have been working on
it while the court case was ongoing)
 Addresses OOS results and instructs
inspectors to be alert
 “Evaluate the company's system to investigate
laboratory test failures. These investigations
represent a key issue in deciding whether a
product may be released or rejected and form
the basis for retesting, and resampling “
 (Most of the information is now in FDA’s guide)
PCI Pharmaceutical Consulting Israel Ltd
13/82
Guide to Inspection: QC Labs





OOS results fall into three categories:
– laboratory error
– non-process related or operator error
– process related or manufacturing process error
Evaluate the company's retesting SOP for compliance with
scientifically sound and appropriate procedures
A very important ruling in one recent court decision sets forth a
procedure to govern the retesting program
This district court ruling provides an excellent guide to use in
evaluating some aspects of a pharmaceutical laboratory, but
should not be considered as law, regulation or binding
legal precedent
The court ruled that a firm should have a predetermined testing
procedure and should consider a point where testing ends and
product is evaluated. If results are not satisfactory, product is
rejected.
PCI Pharmaceutical Consulting Israel Ltd
14/82
After Barr – Current Industry Practice 2007




ALL pharmaceutical companies in developed
countries and many in less developed
countries have SOPs for handling
Out of Specification results
There are still many investigational findings
concerning out of specification results
There are still numerous issues, particularly
with transparency:
What do you report on the COA?
Able laboratories suspended activity in 2005
because of Out of Specification results
PCI Pharmaceutical Consulting Israel Ltd
15/82
When it all goes wrong….Able 2005
The Quality Unit failed to:
 Review computer audit trails in the Waters Empower
Data Acquisition System
 Provide adequate training to analytical chemists
These practices led to:
 The QU releasing batches failing in-process, finish
ed product and stability specifications
 Submission of erroneous data in Annual Reports
and Prior Approval Supplements
 Ceasing manufacture, distribution and recall of all
products as of 13 May 2005 and withdrawal of at
least 5 ANDAs
PCI Pharmaceutical Consulting Israel Ltd
16/82
Resample, Re-injection, Reprocessing
PCI Pharmaceutical Consulting Israel Ltd
17/82
Time for a Break
18
From Able Laboratories 483

Notebooks and binders lacked data from all
testing conducted in the QC Laboratory.
 Records did not include all data such as OOS
results, chromatograms, sample weights, and
processing methods.
 OOS results were substituted by passing results
by Analysts and Supervisors.
 The substitution of data was performed by cutting
and pasting of chromatograms, substituting vials,
changing sample weights and changing
processing methods
PCI Pharmaceutical Consulting Israel Ltd
19/82
That isn’t relevant…it’s fraud

Think about the following scenario
before saying it couldn’t happen:
– You are just starting up your HPLC system in
the morning
– You inject 5 replicates of standard
– The first four give perfect responses
– The fifth is way off / makes no sense,
obviously incorrect

What do you do?
PCI Pharmaceutical Consulting Israel Ltd
20/82
It couldn’t happen in my company



A split second wrong judgment can appear as
fraud to the inspector
You inject one more standard injection and it
gives a perfect response
Clearly you were right and the original “5th”
injection was dirty glassware or mis-injected,
or, or, or……
PCI Pharmaceutical Consulting Israel Ltd
21/82
It couldn’t happen in my company




But you don’t want to show 4 good injections,
one lousy and another good one
So you cut out the bad one and document
only the good ones in the notebook
Sounds “horrendous?”
It has been done, probably in some of your
laboratories! Can you be sure none of your
analysts ever did this?
How can you be so sure?
PCI Pharmaceutical Consulting Israel Ltd
22/82
GMP and OOS Results
Do you know where the term OOS appears in
– EU GMP regulations
it didn’t
as of June 2006 it does
– US GMPs
it doesn’t
– Q7A GMP for APIs
it DOES
because Q7A was written after 1993
PCI Pharmaceutical Consulting Israel Ltd
23/82
Q7A on OOS
PCI Pharmaceutical Consulting Israel Ltd
24/82
Ever heard this….
“The Lab don’t know how
to test”
“Give it to Pete…
he knows how to do
that test…
Dave always gets bad
results!”
PCI Pharmaceutical Consulting Israel Ltd
25/82
OOS Case Study #1






Complex biochemical assay: 6 replicates
Inherent variability allows for wider than usual
specification of 80 – 120%
Results: 45, 50, 46, 52, 65, 69%
What would you think if it happened to you?
The lab technician has 20 years seniority
No other technician is familiar with the test
PCI Pharmaceutical Consulting Israel Ltd
26/82
OOS Case Study #1 - continued
“Must be my mistake!”
 WRONG
 Analyst retested (not in accordance with SOP)
 Results: 72, 69, 81, 80, 82, 81%


NOW what would you think?
PCI Pharmaceutical Consulting Israel Ltd
27/82
OOS Case Study #1 - continued

“Results: 72, 69, 81, 80, 82, 81%

72 and 69% must be “OUTLIERS”

Average 81, 80, 82 and 81 and result passes
 Batch can be released
 Report only this set of results
PCI Pharmaceutical Consulting Israel Ltd
28/82
OOS Case Study #1 - continued





The outcome…..
Product complaints from patients and doctors:
Product is sub-potent
Litigation
Product recall
Investigation reveals weighing error in
production
PCI Pharmaceutical Consulting Israel Ltd
29/82
What is the purpose of testing?

To find out the value of a specific parameter
 To assess if that parameter meets the
pre-determined specification for each lot
 So as to make a sound scientific judgment
regarding product release or rejection
 What happens in your company when there is
an OOS result?
 IDEALLY the analyst doesn’t know the
specification because of “BIAS.”
PCI Pharmaceutical Consulting Israel Ltd
30/82
Is it possible to ensure a correct result?

Not at a 100% certainty level
 Just as it is not possible to prevent an incorrect
result (at the 100% certainty level)
 What is a correct result?
 A result that is identical to the true result….
BUT
 When testing, we NEVER know the true result
PCI Pharmaceutical Consulting Israel Ltd
31/82
Consequences of an Incorrect Test Result?

FALSE NEGATIVE
Product declared fit for use when not
Side Effects
Death

FALSE POSITIVE
Product declared unfit for use when fit
Rejection
Financial Loss
PCI Pharmaceutical Consulting Israel Ltd
32/82
Is it possible to ensure a correct result?

It is possible to ensure a result as close as
possible to the true result
 Using a quality assurance program in the
laboratory
 Validate methods
 Qualify analysts
 Follow methods as written
 Qualify, calibrate and maintain equipment
 Report deviations / malfunctions
PCI Pharmaceutical Consulting Israel Ltd
33/82
Types of Tests





Quantitative: assays and some limits tests
Qualitative: e.g. sterility test, appearance
Chemical
Physical
Microbiological
PCI Pharmaceutical Consulting Israel Ltd
34/82
Chemical Testing




Inherently reliable
Precision is usually considerably better than
for microbiolgical and biochemical testing
Outlier testing is forbidden by the FDA guide
for chemical testing
(usually have less replicates anyway)
Don’t forget case study #1 – How NOT to
handle OOS results
PCI Pharmaceutical Consulting Israel Ltd
35/82
Physical Tests


e.g.1
Black spots in powder
e.g.2
Fill volume
are particularly problematic, since results are
almost certainly correct.
Is there any place for retesting?
In e.g. 1
Probably not
In e.g. 2
Maybe
Is there any place for resampling?
PCI Pharmaceutical Consulting Israel Ltd
36/82
FDA OOS Guidance, October 2006





Draft from September 1998 for comments
Was confusing and open to misinterpretation
particularly by persons not familiar with the
industry
Final guidance is still confusing but a lot better than
the draft
The draft was referenced, as was the Barr court case
during inspections, so the final guidance will definitely
be used
Should be considered in preparing an OOS SOP
PCI Pharmaceutical Consulting Israel Ltd
37/82
FDA Guide on OOS Results





Finalized October 2006
Contains “nonbinding recommendations”
Requires that an investigation is performed
ANYTIME that an OOS is obtained
Wants to include ALL suspect results
Wants an SOP stating number of retests
UPFRONT
PCI Pharmaceutical Consulting Israel Ltd
38/82
Table of Contents



Introduction
Background
Identifying and Assessing OOS Results –
Phase I: Laboratory Investigation
– Responsibility of Analyst
– Responsibilities of Laboratory Supervisor
PCI Pharmaceutical Consulting Israel Ltd
39/82
Table of Contents (continued)

Investigating OOS Results –
Phase II: Full Scale OOS Investigation
– Review of Production
– Additional Laboratory Testing
– Reporting Testing Results

Concluding the Investigation
– Interpretation of Investigation Results
– Cautions
– Field Alert Reports
PCI Pharmaceutical Consulting Israel Ltd
40/82
Out of Specification Results


IF you don’t know the specification you
should never have an OOS, but we don’t
live in an ideal world!
HOW can you have an OOS if:
– the method is validated
– analysts are qualified
– equipment is calibrated and
well-maintained
– notebooks have full record of testing
PCI Pharmaceutical Consulting Israel Ltd
41/82
How to Identify OOS Results


Compare result with specification
Be sure that specification is to required degree
of accuracy and round the result to this value
E.g.

spec 95 - 105
94.9 passes
spec 95.0 - 105.0
94.9 fails
Message for your R&D Department:
think carefully before setting specifications!!!
PCI Pharmaceutical Consulting Israel Ltd
42/82
How to Identify OOT Results

Out of Trend or unusual results are generally
results that:
– MEET the product specification
– ARE outside the control limits of the process,
where control charts are used
or
– ARE different to results usually obtained
(e.g. spec:
95.0 – 110.0
usual results:
98.5 – 101.0
OOT result:
96.4
PCI Pharmaceutical Consulting Israel Ltd
43/82
Initial Assessment of OOS Result

Bear in mind prior:
– Product history
– Process history
– Test history
– Reliability of equipment
– Reliability of the analyst
– Precision of the test (validation)
PCI Pharmaceutical Consulting Israel Ltd
44/82
Out of Specification Results

Failure of a control is NOT an OOS eg:
– Standard shows impurity peak and assays at
incorrect retention time
– Standard shows assay of 80%
– System suitability drift
– In this case all test results SHOULD BE INVALIDATED
and the test REPEATED using an additional aliquot of
the same preparation if possible, or an additional
aliquot from the same sample (What if product degrades
and this is not possible? – need to define
duplicate sampling procedure so that always have
spare material)
– Data collected is retained on file ANYWAY
– What happens if results are OOS? Do you need to err
on the side of caution e.g. possible homogeneity issue?
PCI Pharmaceutical Consulting Israel Ltd
45/82
OOS – Laboratory Investigation


IF you don’t question an “in-spec” result, why
do you question an OOS?
Must have a Retesting Policy:
– Laboratory investigation
•
•
•
•
•
equipment -e.g. glassware, calibration, maintenance
question large differences between replicates
calculations
Product history etc. etc. etc. (to be continued)
4M’s: man, machine, methods, materials
PCI Pharmaceutical Consulting Israel Ltd
46/82
Time for Lunch
47
FDA Guide: Introduction





Applies to chemistry – based testing
[KSG: primarily HPLC and GC]
Includes in-process testing except (footnote),
where the purpose is to prevent process drift
[KSG: discuss – could still have OOS!]
Principles apply to CONTRACT firms
Timely, unbiased investigation
[KSG: OOS Policy is major SOP approved by
senior management]
PCI Pharmaceutical Consulting Israel Ltd
48/82
FDA: Laboratory Investigation

The source of OOS should be identified as:
– Aberration of measurement process
– Aberration of manufacturing process
[KSG: sampling process?]

Even if the batch is rejected, an investigation is
required:
– Other batches involved?
– Other products involved?

Need written investigation, conclusions, follow-up
PCI Pharmaceutical Consulting Israel Ltd
49/82
FDA: Laboratory Investigation

Assess accuracy of lab data:
– Before test preparations are discarded
(composite / homogeneous source of aliquot tested)
– Hypothesis testing using same test preparation
for laboratory error or instrument malfunction
– IF no meaningful errors were made [found?],
conduct a full scale OOS investigation
– Contract lab conveys data and findings to you
PCI Pharmaceutical Consulting Israel Ltd
50/82
OOS Flow Chart
OOS result
Laboratory Investigation
Report to QA (copy in batch file?)
Convincing evidence
of
Laboratory Error
Inconclusive
Invalidate result
Perform new test on same sample
QA Investigation
Production Error
Sampling Error
Inconclusive
No evidence of production error
Correct if possible
or
Reject Batch
Resample
Double sample
Revise sampling procedures
Retest
??? further aliquots
from original sample
All within
specifications
One result
OOS
QA decision
regarding
batch disposition
Reject Batch
PCI Pharmaceutical Consulting Israel Ltd
51/82
Responsibility of Analyst

To follow test procedure as written
 To be alert to errors and STOP test BEFORE o
btaining the result if error is suspected,
recording what happened e.g. spill
 Analyst responsible for ensuring that
instruments meet performance specifications
and are properly calibrated [KSG: maintained?]
 Once an OOS result is obtained to review all
records relative to the test to identify possible
laboratory error
PCI Pharmaceutical Consulting Israel Ltd
52/82
Supervisory Role

Supervisors / team leaders / laboratory head:
– Should be experienced analysts
– Frequently audit while tests ARE BEING
performed
in order to be able to objectively investigate
OOS results
PCI Pharmaceutical Consulting Israel Ltd
53/82
Responsibility of Supervisor


Objective assessment without
preconceived assumptions as to cause of OO
S
Immediate assessment may include:
– Re-examination of:
• actual solutions
• Test units
• Glassware
used in the original measurements and preparations
This could provide more credibility for laboratory
error hypotheses
PCI Pharmaceutical Consulting Israel Ltd
54/82
Responsibility of Supervisor
To INVESTIGATE:
– Review notebook / worksheet with analyst :
• Was method was followed: with a copy of the
method in your hand, have the analyst describe
exactly how they performed each step:
confirm that the method was understood & followed
• Review raw data: Perform calculations again
including checking dilution schemes
• Unauthorised changes to automated calculations
• Examine reagents, (reference) standards, solutions
• Examine glassware
• Performance of instruments
PCI Pharmaceutical Consulting Israel Ltd
55/82
Hypothesis Testing

If you re-create an OOS conclusions are stronger
 Might retest recent sample to confirm / refute
equipment malfunction: retro consequences?
 Examine retained samples e.g.
– Re-inject solutions where a transient equipment
malfunction is suspected [air bubbles]
Hard to prove, but re-injection of same
preparation can provide strong evidence
– Release rate testing of slow release tab / cap:
retest of original unit may show it was damaged during
testing
– Additional extraction to show that problem is
method related
– then revise method
56/82
PCI Pharmaceutical Consulting Israel Ltd
Hypothesis Testing, Case Study #2





LAL test on radioactive product (short half-life)
is positive
Product history – no previous failure
Other products tested in same series passed
Initial laboratory investigation: no evidence
of lab error
Dilution used: 1:70
specification allows up to 1:140 dilution
PCI Pharmaceutical Consulting Israel Ltd
57/82
Hypothesis Testing, Case Study #2
– Perform retest in parallel with production
investigation because of short half-life
– Production investigation included sampling
glassware and equipment for LAL residues
NO positive results
– Repeat test at 1:70 and 1:140 dilutions
Results were in spec i.e. no positive LAL
– New LAL reagent prepared and same sample
tested with old and new reagent:
• Old: failed
• New: passed
PCI Pharmaceutical Consulting Israel Ltd
58/82
Out of Specification Results

If the laboratory investigation is conclusive
–
–
–
–
Document findings
INVALIDATE original test
Perform NEW test on same sample
Report original result with investigation as well as new
result in batch record for QA review prior to release
– COA carries new result only; some companies use an
asterisk and indicate that there was an OOS

If the laboratory investigation is NOT conclusive
inform QA (or customer for contract lab)
PCI Pharmaceutical Consulting Israel Ltd
59/82
Out of Trend Results



If the laboratory investigation is conclusive or
inconclusive, consult with QA
In most cases, DO NOT perform any additional
testing or sampling
Make product disposition judgment based on:
– Original result
– Product history (e.g. stability data – statistical analyses
of particular use here)
– Batch history (e.g. review indicates that there were no
processing errors / there were errors)
– Other investigational findings
PCI Pharmaceutical Consulting Israel Ltd
60/82
And now some work for you.....
Specification is 90.0 – 110.0% for an oral
suspension
Three batches are tested simultaneously, each
sample in duplicate
Two batches show (average) 98.2% and 99.7%
respectively
Third batch gives one result of 88.2% and a
second result of 89.3%
PREPARE A CHECKLIST OF QUESTIONS
TO REVIEW AS PART OF THE
LABORATORY INVESTIGATION
PCI Pharmaceutical Consulting Israel Ltd
61/82
FDA: Full-Scale OOS Investigation








Use a pre-defined procedure
Production/ process review and / or additional
laboratory work
Identify root cause and implement CAPA
QA / QCU responsibility, includes CMOs if used
Documented in the batch record
Involves all aspects of manufacture, quality
control and sampling
Describes corrective actions and endpoint
Is performed PRIOR to ANY retesting
PCI Pharmaceutical Consulting Israel Ltd
62/82
FDA: Full-Scale OOS Investigation

If cause of OOS is identified, batch is rejected
 In this case need CAPA on process / product
 May not identify cause and may need additional
lab testing:
– Retest additional portion of original sample
– Resample
PCI Pharmaceutical Consulting Israel Ltd
63/82
FDA: Retesting and Resampling





Use another analyst? Where possible
The maximum number of retests should be
specified in advance in an SOP
May, on rare occasions, deviate from SOP but
with documented rationale and protocol
The number may vary depending upon the
variability of the test method and NOT
depending on the results obtained
Resampling raises questions as to sampling
procedure validity
PCI Pharmaceutical Consulting Israel Ltd
64/82
Interpretation and Results Reporting


Report all results, passing and suspect and
consider for batch release decision
Averaging
– Has appropriate and inappropriate uses:
• Appropriate: optical rotation test: several discrete
measurements averaged for result; microbiology;
HPLC average of 2 or 3 peak responses from
replicate injections of the same preparation = one
test and one result as described in the test method
Have acceptance criteria for deviation between replicates
This is different to analysis of different portions of
a single batch
PCI Pharmaceutical Consulting Israel Ltd
65/82
Interpretation and Results Reporting

Averaging
– Inappropriate uses:
• May conceal variations in different portions of the
batch or sample
e.g. for powder blend uniformity or dosage form
uniformity of content.
• Averaging the results of the original OOS and
additional retest or resample results is inappropriate
• All individual results must be provided and
considered by the QCU / QA for release
PCI Pharmaceutical Consulting Israel Ltd
66/82
Time for a Break
67
Outlier Tests

Possible use of the outlier test should be
determined in advance
 Should be written in an SOP for data
interpretation
 Should include the specific test to be used
 Should specify the minimum number of results
required to obtain a statistically significant
assessment from the test
 For validated chemical tests its use is suspect
PCI Pharmaceutical Consulting Israel Ltd
68/82
FDA: Concluding the Investigation

Evaluate results and determine batch quality
 Release decision by QCU / QA
 An initial OOS does not necessarily mean
that the batch fails and must be rejected.
 Where the suspect result is invalidated, the
result should not be used to evaluate the
quality of the batch or lot
 For inconclusive results – give full
consideration to the OOS result
PCI Pharmaceutical Consulting Israel Ltd
69/82
FDA: Concluding the Investigation



Example given shows seven retest results
which gives the only indication in the guide
regarding numbers of retests
The example given is also extreme:
89.5% OOS
99.0, 98.9, 99.0, 99.1, 98.8, 99.1, 99.0%
Consider method precision and validation data
in making release / reject decision
PCI Pharmaceutical Consulting Israel Ltd
70/82
FDA: Cautions



Where a series of assay results (to produce
a single reportable result) have some
individual results OOS and some in spec and
all within the known method variability,
passing results no more likely to represent
the true value than the OOS result.
In this case the company must err on the side
of caution and reject the batch
A result that is low but in specification should
also be a cause of concern
PCI Pharmaceutical Consulting Israel Ltd
71/82
FDA on Field Alert Reports (FARs)



For products with approved (a)/NDAs,
regulations require a FAR within 3 working
days concerning any failure of a distributed
batch to meet any of the specifications
established in an application
Unless the OOS result is found to be invalid
within 3 days, an initial FAR should be
submitted
A follow-up FAR should be submitted when
the investigation is completed
PCI Pharmaceutical Consulting Israel Ltd
72/82
OOS Case Study #3

BN 200602 gave an assay result for 3 month
stability study: 89.2% (long term / RT)
 Limits: 90.0 – 110.0%
 Test performed at a contract laboratory using
an internal instrument control
 BN 200701 tested at the same time gave a
result of 95.4% (i.e. in spec. – release test)
DO YOU INFORM THE REGULATORS?
PCI Pharmaceutical Consulting Israel Ltd
73/82
OOS Case Study #3 – cont/




The samples have to be diluted during
preparation and as far as records showed
there was no evidence of laboratory error
Calculations were satisfactory
No evidence that reagents were outdated or
faulty technique
Control sample showed a 3.5% difference at
beginning and end of run: usually around 1%;
NMT 5% allowed by method
PCI Pharmaceutical Consulting Israel Ltd
74/82
OOS Case Study #3 – cont/



Batch number 200701; in-process result: 98.2%
Batch number 200602; time zero result: 92.8%
i.e. an apparent difference of 3.5% between
results in both cases
Hypothesis formulated:
problem is with the control sample / equipment
and this can be tested using both batches
i.e. in-spec and OOS
PCI Pharmaceutical Consulting Israel Ltd
75/82
OOS Case Study #3 – cont/

Retesting Protocol (before start of retesting):
– BN 200602: OOS on inverted stability sample
at 3 months
Perform retest at three dilutions in duplicate
on:
• Same sample
• Upright sample
= 6 results vs 2 original
= 6 additional results
– BN 200701: In spec. but 3.5% lower than IPC
Perform retest at three dilutions in duplicate
• Same sample
= 6 results vs 2 original
PCI Pharmaceutical Consulting Israel Ltd
76/82
OOS Case Study #3 – cont/

Outcome:
– BN 200602: All retest results in specification
average result: 92.2%
– BN 200701: All retest results in specification
average result: 97.9%
– Control sample: difference of 1% beginning and
end of run
PCI Pharmaceutical Consulting Israel Ltd
77/82
OOS Case Study #3 – Conclusion



Original OOS result for BN 200602 is invalid
Original OOT result for BN 200701 is invalid ?
CAPA required regarding external laboratory
as follows:
– Tighten allowed limits for control sample
(NMT 5% is too high)
– Revalidate method? Investigation showed
validation last done 15 years ago and a new
instrument had been introduced since then!
– Closely follow BN 200602 at additional stability
stations
78/82
PCI Pharmaceutical Consulting Israel Ltd
OOS Logs and Trending Data


Should keep a record of OOS results
Analyse periodically according to:
– Analyst
– Instrument
– Product
– Test method

Take appropriate actions where repeat
problems appear
PCI Pharmaceutical Consulting Israel Ltd
79/82
And in Conclusion....

Always report OOS or OOT result to
Supervisor
 Have a pre-approved SOP that provides a
flow-chart for handling OOS
 Conduct detailed and genuine investigation
 Document investigational findings clearly and
concisely
 Transparency regarding reporting on COA
 At some point...notification of regulators
PCI Pharmaceutical Consulting Israel Ltd
80/82
In conclusion







All test results are subject to some element of doubt
Use controls and validated, approved test methods
Use replicates where inherent variation exists
Use qualified, calibrated and well maintained
equipment
Perform trend analyses and report deviations
NEVER continue a suspect test
At some point, testing ends and a product disposition
decision must be made
PCI Pharmaceutical Consulting Israel Ltd
81/82
Thank you for your attention
Any questions?
Find me at [email protected]
PCI Pharmaceutical Consulting Israel Ltd
82/82